Assembly Language for Beginners

1.18. REPLACING ARITHMETIC INSTRUCTIONS TO OTHER ONES

; File c:\polygon\c\2.c
mov eax, DWORD PTR _a$[esp-4]
add eax, eax
add eax, eax
add eax, eax
ret 0
_f ENDP
_TEXT ENDS
END

Multiplication using shifting

Multiplication and division instructions by a numbers that’s a power of 2 are often replaced by shift in-
structions.

unsigned int f(unsigned int a)
{
return a*4;
};

Listing 1.189: Non-optimizing MSVC 2010

_a$ = 8 ; size = 4
_f PROC
push ebp
mov ebp, esp
mov eax, DWORD PTR _a$[ebp]
shl eax, 2
pop ebp
ret 0
_f ENDP

Multiplication by 4 is just shifting the number to the left by 2 bits and inserting 2 zero bits at the right (as
the last two bits). It is just like multiplying 3 by 100 —we just have to add two zeros at the right.

That’s how the shift left instruction works:

7 6 5 4 3 2 1 0

CF 7 6 5 4 3 2 1 0 0

The added bits at right are always zeros.

Multiplication by 4 in ARM:

Listing 1.190: Non-optimizing Keil 6/2013 (ARM mode)

f PROC
LSL r0,r0,#2
BX lr
ENDP

Multiplication by 4 in MIPS:

Listing 1.191: Optimizing GCC 4.4.5 (IDA)

jr $ra

sll $v0, $a0, 2 ; branch delay slot

SLLis “Shift Left Logical”.

Multiplication using shifting, subtracting, and adding

It’s still possible to get rid of the multiplication operation when you multiply by numbers like 7 or 17 again
by using shifting. The mathematics used here is relatively easy.